hook-creator

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Provides a Node.js template and instructions for executing external binaries via execFileSync. It correctly advises against using execSync with shell interpolation to prevent command injection vulnerabilities.\n- [PROMPT_INJECTION]: The skill documents the creation of hooks that ingest untrusted tool inputs to inject additional context into the AI agent's session. This creates a surface for indirect prompt injection where malicious data in processed files could influence agent behavior.\n
  • Ingestion points: Hook scripts read JSON data containing tool inputs from stdin (SKILL.md).\n
  • Boundary markers: The template does not explicitly demonstrate the use of delimiters for the injected additionalContext.\n
  • Capability inventory: Uses execFileSync for subprocess calls (SKILL.md).\n
  • Sanitization: The skill provides a 'Security Rules' section recommending path traversal checks and exclusion of sensitive files (e.g., .env, .pem).
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 08:41 AM